LigthsON Voice Agent
v2.0 • RAG + NLU + VOICE ENGINEReal-time Conversational Voice Finance Assistant with Automated Firestore CRUD
Connecting to API...
Voice Engine Idle
🎙️ Conversational Voice Assistant
Talk naturally with your AI personal finance agent. Real-time audio streams while NLU, RAG, and Firestore execute database CRUD operations seamlessly.
📡 Real-Time WebRTC Audio
🧠 RAG Vector Memory
🔍 NLU Entity Parsing
⚡ Real-Time Firestore CRUD
🤖
Click "Talk to Your Voice Assistant" to join voice room and start talking.
Live Voice Persona & Language:
👨💼 Ravi (Deep Male) • 🌐 Auto Detect
Voice:
Language:
💬 Live Conversation Feed
0 Turns🤖
LigthsON Assistant
Just now
Hey! I'm your personal finance assistant. 🎙️ Tap the mic and say something like "Spent 350 on groceries", or just type below. I'll log it, query it, update it — whatever you need!
💳 Live Firestore Transactions Ledger
Auto-SyncedLoading transactions ledger...
🧠 RAG Vector Memory & Financial Context
RAG ActiveTransactions are embedded in the semantic vector space in real-time. The agent queries past financial history to provide deep conversational answers.
Speak or query the assistant to inspect retrieved vector embeddings and LLM analysis.
📜 Section 3.2 Schema Payload
Confidence: --// Latest NLU and CRUD payload will be rendered here🔌 Mobile & Backend Integration Guide
Connect your React Native app or backend to this hosted Voice Agent service:
POST
/api/voice/token
Issue WebRTC voice participant token for real-time streaming
POST
/api/voice/agent/conversation
Conversational endpoint executing NLU + RAG + Firestore CRUD in one turn
GET
/api/voice/transactions
Fetch all real-time Firestore user transactions (Expenses, Income, Investments)
POST
/api/voice/process-text
Process raw natural language text utterance or multi-turn conversational session